A flexible real-time surface electromyography-based wheelchair control system using upper arm or neck muscles.

OBJECTIVE

People with movement disabilities face difficulties in independent mobility. The main objective of this study was to develop and evaluate a real-time electro-myography (EMG) based wheelchair control framework with upper arm or neck muscles and to investigate feasibility for assistive mobility applications.

METHODS

A feature optimisation framework was employed to extract the most significant features for movement recognition. The proposed model was trained and tested in both offline and real-time control. Feasibility was assessed through classification accuracy, repeatability measures, response latency and preliminary validation involving two individuals with disabilities.

RESULTS

With subject-independent methodology, wheelchair activation with upper arm muscles achieved an accuracy of 95% in real time, whereas for neck muscles, it was 98.6% in able-bodied participants. For neck muscles, two participants with disabilities, namely a transhumeral amputee and a mobility-impaired participant because of poliomyelitis were evaluated and achieved an accuracy of 100%. The average system response latency for upper arm movements was found to be 430 ms for able-bodied participants, whereas for neck muscles, it was 360 ms for able-bodied participants and 400 ms for participants with disabilities.

CONCLUSION

The proposed framework demonstrated the feasibility of real-time EMG-based wheelchair control in a controlled environment. However, further studies involving larger and more diverse participant cohorts are needed to evaluate its generalisability and broader clinical applicability.